Scalability & infrastructure
The same platform runs 5-driver and 500-driver fleets.
No driver, user or admin caps on any plan. Plans only control monthly trip quota; the architecture is identical across Basic, Professional and Enterprise.
+87
Active operators
14
LATAM countries
+270
Cities
+970
Published apps
Stack running in production today
- Next.js 16 + Vercel — Dashboard with edge autoscaling.
- PostgreSQL on Neon — Autoscaling pool + branching for staging.
- Centrifugo on Hetzner — Realtime broadcast-then-fetch for dispatch + tracking.
- Native Flutter — Android + iOS from a single codebase, builds in minutes.
- Cloudflare R2 + Vercel Edge — Global CDN for static assets and user files.
- OAuth 2.0 + PKCE + JWKS — Auth ready for AI agents with Dynamic Client Registration.
Capacity & growth
- NO driver, user or admin caps on any plan.
- Plans (300/1000/3500 monthly trips) only control the quota — exceeding triggers a plan upgrade, not an outage.
- Operators with 100+ drivers in production since 2024 without architecture migration.
- Multi-country tenants with per-zone currency + timezone.
- Android + iOS builds in parallel on dedicated servers (Mac mini + Linux VPS).
- 99.9% uptime target on the realtime layer.
Why LATAM-first
Cabgo is built specifically for Latin America: Spanish docs, integrations with local payment methods (SPEI, OXXO, Mercado Pago, Conekta), WhatsApp Business as primary support channel for the operator and riders, timezones and currencies configurable per city. Active operators today cover Mexico, Colombia, Peru, Ecuador, Chile, Argentina, Bolivia, Paraguay, Uruguay, Venezuela, Guatemala, El Salvador, Honduras, Costa Rica and Dominican Republic.
Scale and infrastructure: frequent questions
- Production operators already process more than 5,000 trips per day in a single tenant without architecture migration. The realtime layer (Centrifugo) and the PostgreSQL database scale horizontally, and Android/iOS builds run on dedicated pools (Mac mini + Linux VPS). The practical per-tenant ceiling is far above current volumes.
- No. Every plan (Basic, Professional, Enterprise) ships with UNLIMITED drivers, riders, and admins. Pricing scales only with total monthly trip volume. A fleet of 5 drivers and one of 500 run on the same platform with no caps.
- Isolation by companyId at both the application and database layer. Every rider, driver, and dashboard query includes the companyId filter; endpoints validate that the bearer token belongs to the tenant. We do not share driver pools, rider data, or reports between operators.
- Target 99.9% uptime on the realtime layer. Infrastructure runs on Vercel (Next.js + functions), Neon (PostgreSQL), Cloudflare R2 (assets), Centrifugo (realtime, dedicated VPS), and build servers (Mac mini + Linux). Public status at cabgo.app/status.
- Yes. Today there are active operators across 17 Latin American countries simultaneously. Each tenant configures its timezone, currency, language, and zones; everything else adapts automatically (date formats, currency display, per-km vs per-mile tariffs, local gateways).
- Vercel: the mobile apps have circuit breakers and offline caching for critical operations (trips in progress continue; new requests queue). Maps: if HERE fails momentarily, you can switch to Google Maps at the tenant level without re-deploying the app. Realtime runs on a dedicated VPS independent of Vercel.